ndclab / readaloud-valence-dataset Goto Github PK

Seek volunteers for full run-through of REDCap and Pavlovia.

1. Recheck all switch group formulae.
2. Clean up switch tab in main Excel and add public-facing note.

• make code public
• add link in GH repo

Add passage characteristics to data dictionary.

Reminder to specific include code to check the practice rounds as all participants will complete the DCCS task, even if isForward == 3.

pos-neg

1. dolphins ✅ (question: 👍)
2. dams ✅ (question: 😦)
3. congo ✅ (question: 👍)
4. antarctica ✅ (question: 😦)
5. broccoli ✅ (question: 👍)
6. bats ✅ (question: 😦)
7. ice-fishing ✅ (question: 👍)
8. flying ✅ (question: 😦)
9. realty ✅ (question: 👍)
10. skunkowl ✅ (question: 😦)

neg-pos

1. mantis ✅ (question: 👍)
2. depression ✅ (question: 😦)
3. bees ✅ (question: 👍)
4. sun ✅ (question: 😦)
5. vegas ✅ (question: 👍)
6. cars ✅ (question: 😦)
7. caramel ✅ (question: 👍)
8. dogshow ✅ (question: 😦)
9. grizzly ✅ (question: 👍)
10. dentist ✅ (question: 😦)

Add dates and location of data collection, plus demographics table, to readme file.

Include extended Kousta valence ratings in Excel for LDT stimuli.

• main passage task
• practice + main LDT task
• add mic test ❌
• add mic input ❌
• add practice feedback
• add passage counter
• break loop for between passage break
• standardize stimuli text to 'cm'
• counterbalance two word lists: pos>neg >> neg>pos >> and neg>pos >> pos>neg>>
• ensure participant cannot onset at zero seconds (for preprocessing script)
• add DCCS to experiment
• add Zoom setup instructions
• add file transfer instructions

• passage/word/syllable file for each stimulus passage
• recalculate syllable counts in main Excel

Overview

• LDT: prior authors gave me everything, pulled together semi-comprehensive list of interesting (sub)lexical characteristics for future analysis efforts
• readAloud: Drafted four passages (2 x neg>pos, 2 x pos>neg), with efforts to control valence using Warriner et al. list. Once all passages complete, will pull (sub)lexical characteristics like those for LDT plus passage-level characteristics.
• PsychoPy: downloaded RWE readAloud task and began making modifications.
• Preprocessing: scripted initial code (for mantis text) to convert human-coded Excel (by syllable) into usable variables.

GitHub

Check readme and contributing for merge.

Design Questions

• per passage reading: 1.5 minutes, plus challenge question -- how many passages?
• prediction 1a: do we need to go full factorial?
• blocks: should we create blocks with breaks or just run them each back-to-back (with participant taking short breaks between, as they wish? (correlated questions: need for any counterbalancing? how to break loop after last passage and challenge question?)
• challenge questions: discuss difficulty level and range of levels (some harder, some easier?), discuss whether to provide feedback?

PsychoPy Questions

• Question on Layout>Spatial Units for text fields: "from exp settings"?
• LDT: Kousta et al. used F/J for keyboard input. Choose between this and Z/M.
• Not rendering change in screen color?